Forecasting in the age of big data

A recent Gartner survey demonstrated just how important digital business has become. Some 62 percent of respondents said they have a management initiative or transformation plan in place to ensure that their businesses are becoming more digital, and another 47 percent are challenged by their board of directors to do so.

Finance teams aren’t spared the drive to digitize. According to a Workday Global Leader Finance survey, digital disruption is “front of mind” for the next generation of finance executives.

What exactly is driving these digital transformation initiatives now underway at virtually every Fortune 500 company? In every way, business has become digitized, and that in turn means organizations are dealing with new and massive datasets that contain important insights about their customers, competition and emerging trends.

No one needs to harness that data more than the finance team; you’re the one who is tasked with forecasting a future for your company, and monitoring whether or not you’re on track, or whether pivots must be made to exploit unexpected opportunities or avert disasters. These are tall orders.

And yet, as Ashish Pareek pointed out in CFO magazine, when it comes to forecasting, too many financial teams are “ignoring real-time metrics in favor of historical data.” He suggests that the reliance on historical data may be a direct result of the overwhelming nature of big data and the uncertainty of which metrics best inform forecasting projections.

These aren’t idle musings on the part of Mr. Pareek. It’s easy to be daunted by the technology one needs to implement in order to capture, validate and "munge" (or wrangle) all that data.

Fortunately, finance teams have a lot of robust resources available to them today. Consider the detailed and nuanced data contained in your customer relationship management, order management and other systems. Every day your sales teams will enter sales and potential sales into your CRM system, detailing the deal size, terms, market, customer type, product type and so on. Meanwhile, your systems track every nickel your company will pay in compensation and benefits for your full-time, part-time and contract employees. These systems, which are market-tested, provide a level of detail and accuracy you would need an army to create manually. Apply that to your forecasts, and you’re in a much better position to predict and/or pivot based on market realities.

If you’re still using a spreadsheet to build your forecast, then you are absolutely limiting your ability to use anything other than historical data. Take a cue from your software development team — do you think they rely on a spreadsheet to track all of the tasks and dependencies needed to get a release out the door?

Like the development team, finance should adopt a data-driven project management approach to forecasting, one in which all outcomes — revenue earned, expenses incurred — are tied to every dependency that affects them automatically. You can then link your platform to your general ledger, so anytime marketing hires a vendor, HR hires a new admin, or operations signs a new lease, you can see the impact of those actions on your forecast in real time.

To be sure, the data that’s siloed in your system is just a tiny fraction of the data revolution. In the future, finance teams will want to incorporate a wider set of second- and third-party data, such as global economic trends. But in the meantime, this level of automation will help you put all of these datasets to use in creating a realistic plan, which is essentially your forecast. Things can and will change, but by leveraging your data, you can have a lot more confidence in your forecasts.